Abstract: Railway electronic payment platform carries railway passenger and freight electronic payment and capital settlement business. With the rapid development of passenger and freight business, the amount of transaction data increases rapidly, and the problem of low efficiency of payment check processing becomes increasingly prominent. Combined with the process of payment check, a data processing scheme based on distributed parallel computing is proposed, in which the messaging middleware Kafka is used to collect data, Hadoop and Spark are used to build the big data processing and multi-task parallel computing environment of payment check processing. Based on distributed query engine, it provides efficient and flexible query interfaces of payment check results. It has been verified in test that the data processing efficiency of the railway electronic payment platform's payment checking has been satisfactorily improved and the scalability of the data processing platform has also been improved through the technical upgrading and transformation, which provides a guarantee for the railway electronic payment platform to better support the development of railway business.